Tmc gene therapy restores auditory function in deaf mice

D E A F N E S S
Tmc gene therapy restores auditory function in deaf mice
Charles Askew,1,2
Cylia Rochat,3
Bifeng Pan,1
Yukako Asai,1
Hena Ahmed,1
Erin Child,1
Bernard L. Schneider,3
Patrick Aebischer,3
Jeffrey R. Holt1
*
Genetic hearing loss accounts for up to 50% of prelingual deafness worldwide, yet there are no biologic treatments
currently available. To investigate gene therapy as a potential biologic strategy for restoration of auditory function in
patients with genetic hearing loss, we tested a gene augmentation approach in mouse models of genetic deafness. We
focused on DFNB7/11 and DFNA36, which are autosomal recessive and dominant deafnesses, respectively, caused by
mutations in transmembrane channel–like 1 (TMC1). Mice that carry targeted deletion of Tmc1 or a dominant Tmc1 point
mutation, known as Beethoven, are good models for human DFNB7/11 and DFNA36. We screened several adeno-
associated viral (AAV) serotypes and promoters and identified AAV2/1 and the chicken b-actin (Cba) promoter as an
efficient combination for driving the expression of exogenous Tmc1 in inner hair cells in vivo. Exogenous Tmc1 or its
closely related ortholog, Tmc2, were capable of restoring sensory transduction, auditory brainstem responses, and
acoustic startle reflexes in otherwise deaf mice, suggesting that gene augmentation with Tmc1 or Tmc2 is well suited for
further development as a strategy for restoration of auditory function in deaf patients who carry TMC1 mutations.
INTRODUCTION
Hearing loss is the most common sensory deficit in the world, with
both genetic and environmental factors causing dysfunction of the pri-
mary sensory cells of the inner ear, known as hair cells (1). Hair cells
convert mechanical stimuli into electrical signals and are essential for
normal auditory and balance functions. Unfortunately, hair cells lack
the ability to regenerate; thus, hair cell damage or death is cumulative,
causing progressive hearing loss. The current standards of care for
hearing loss are hearing aids or cochlear implants, which provide in-
complete restoration of function in a limited patient population. Phar-
macologic, stem cell, and gene therapies are being explored as alternative
therapies (1). Of these possible strategies, gene therapy may be best
suited for restoration of hair cell function in genetic hearing loss
(1–4). However, few studies have provided proof-of-principle evidence
supporting gene therapy as a viable strategy for restoration of auditory
function in mouse models of genetic hearing loss. One notable excep-
tion is the restoration of auditory function in mice lacking vesicular
glutamate transporter 3 (VGLUT3), a glutamate transport protein ex-
pressed in auditory inner hair cells (IHCs), required for synaptic trans-
mission from IHC to postsynaptic neurons of the 8th cranial nerve (5).
The authors of that study used adeno-associated viral (AAV) vectors
to deliver the coding sequence for VGLUT3 into IHCs of early postnatal
Vglut3 knockout mice. Although an important advancement, VGLUT3
mutations are not common in humans and, when present, are dom-
inant, suggesting that the clinical utility of VGLUT3 augmentation
may be limited.
To explore gene therapy for a common form of genetic hearing
loss that affects hair cells, we used mice that carry mutations in trans-
membrane channel–like gene 1 (Tmc1). Mutations in human TMC1
account for 4 to 8% of genetic deafness in some populations (6, 7). To
date, 40 TMC1 mutations have been identified that cause deafness in
humans (8, 9). Most are recessive and cause prelingual deafness, whereas
at least five are dominant and cause progressive hearing loss with on-
set during the mid-teen years (7), suggesting possible windows of op-
portunity for clinical intervention.
Although the precise molecular function of TMC1 is unclear, there
is agreement that TMC1 and its closely related ortholog, TMC2, affect
the permeation properties of sensory transduction channels in auditory
hair cells (10–12) and are likely channel components (11). Mice defi-
cient in Tmc1 and Tmc2 lack sensory transduction, are deaf, and suffer
severe balance dysfunction despite the presence of normal hair cell
morphology (10) and hair cells that survive into mature stages (13).
Mice that carry the Beethoven (Bth) (14) point mutation p.M412K in
TMC1 retain sensory transduction but have reduced calcium perme-
ability (11). Beethoven mice are an excellent model for dominant-
progressive hearing loss (DFNA36) in humans who carry an identical
substitution in the orthologous position (p.M418K) of the human
TMC1 gene (15). Mice that carry Tmc1 deletions (10) are good models
for recessive hearing loss (DFNB7/11) in humans with loss-of-function
mutations in TMC1.
Previously, adenoviral vectors were used in vitro to introduce the
coding sequence for Tmc1 or Tmc2 into hair cells excised from mice
deficient in Tmc1 and Tmc2 (10). These experiments demonstrated
partial restoration of sensory transduction in cultured hair cells in vitro.
To extend these studies to an in vivo setting and to develop gene ther-
apy strategies to treat genetic deafness in humans, we designed AAV
vectors that carried the coding sequence for Tmc1 or Tmc2 and in-
jected them in the ears of Tmc1 mutant mice. Here, we demonstrate
that Tmc1 and Tmc2 are functionally redundant, and that either gene
can restore sensory transduction and partial auditory function in vivo
in mice that carry recessive Tmc1 mutations. In addition, we used
Tmc2 gene therapy to preserve auditory function and hair cell survival
in mice that carried dominant Bth mutations in Tmc1. Our results
support continued development of gene therapy strategies for hearing
restoration in humans with genetic deafness.
RESULTS
AAV2/1 targets cochlear hair cells in vitro
To identify AAV serotypes with the highest viral transduction rate in
cochlear hair cells, we incubated AAV-Cmv-eGFP reporter vectors
1
Department of Otolaryngology, F.M. Kirby Neurobiology Center, Boston Children’s
Hospital and Harvard Medical School, Boston, MA 02115, USA. 2
Neuroscience Graduate
Program, University of Virginia, Charlottesville, VA 22908, USA. 3
Brain Mind Institute, École
Polytechnique Fédérale de Lausanne, Lausanne, CH-1015 Lausanne, Switzerland.
*Corresponding author. E-mail: jeffrey.holt@childrens.harvard.edu
R E S E A R C H A R T I C L E
www.ScienceTranslationalMedicine.org 8 July 2015 Vol 7 Issue 295 295ra108 1
onJuly9,2015Downloadedfrom

containing capsid serotypes 1, 2, 6, 8, or 9, at titers that ranged from
3 × 1010
to 4 × 1013
genome copies (gc)/ml, with organotypic mouse
cochlear cultures. Green fluorescent protein (GFP)–positive hair cells
were evident in all cultures. Confocal images from the mid-basal region
of the organ of Corti demonstrated viral transduction of hair cells for
each serotype tested at an effective viral concentration of 3 × 1010
to
3 × 1011
gc/ml (Fig. 1A). The total num-
ber of hair cells per cochlea ranged from
1575 to 3046, depending on the quality of
the dissection, with an average of 2348 ±
389 (±SD; n = 28) (Fig. 1B). Quantifica-
tion of viral transduction rates for whole
cochleas revealed that AAV serotype 2/1
transduced the greatest number of hair
cells at equivalent viral titers for each se-
rotype. AAV2/1 transduced an average of
58% hair cells along the length of the co-
chlea, compared with 14% for AAV2/6, the
serotype with the next highest viral trans-
duction rate (Fig. 1B).
We noted a tonotopic gradient for vi-
ral transduction, apparent for AAV2/1 at
all concentrations tested (Fig. 1C), with
more total hair cells expressing GFP in
the base of the cochlea (up to 95%) than
at the apex (up to 54%). The rate of viral
transduction of IHCs declined sharply
from base to apex (from 81 to 5%; n = 7),
whereas viral transduction rates in outer
hair cells (OHCs) persisted at higher rates
along the base-to-apex axis (from 84 to
57%; n = 7). The mechanism of the basal-
apical gradient is not clear.
Exogenous promoters drive
expression in cochlear cultures
Next, we examined the activity of differ-
ent promoters in cochlear cultures in vitro
using the AAV2/1 vector for delivery and
enhanced GFP (eGFP) expression as a
readout of promoter activity. Promoters
were chosen from three different sources
that are known to have constitutive activ-
ity in most cells types: cytomegalovirus
(Cmv), chicken b-actin (Cba), and mouse
phosphoglycerate kinase 1 (Pgk1). Addi-
tionally, we investigated the activity of the
synapsin 1 (Syn1) promoter, which is known
to be active in cells with synaptic machin-
ery but has not been investigated in hair
cells. We found that both Cmv and Cba
promoters drove robust eGFP expression
in hair cells, as well as many types of sup-
porting cells in the cochlea (Fig. 1D).
Surprisingly, although phosphoglycerate
kinase is an enzyme present in most cells,
the Pgk1 promoter drove eGFP expression
only in supporting cells of the inner sulcus
(Fig. 1D). We also observed Syn1-driven eGFP expression in spiral gan-
glion neurons (Fig. 1D), consistent with a recent report in mice (16) and
the localization of synapsin protein (17). There was no detectable eGFP
expression in IHCs or OHCs despite the presence of ribbon synapses
in these cells. Because both Cmv and Cba drove robust exogenous
gene expression in hair cells, these promoters were chosen for further
Fig. 1. Screen for AAV serotype and promoter in cochlear hair cells. (A) Representative confocal
images of the mid-base of cochlear cultures exposed to AAV-Cmv-eGFP with capsid serotypes indicated.
Wild-type (WT) cochleas were dissected at P0 and exposed to viral concentrations of 3.3 × 1010
gc/ml
(AAV2/1, AAV2/2, and AAV2/6) or 3.3 × 1011
gc/ml (AAV2/8 and AAV2/9) for 24 hours. The tissue was
cultured for 7 days, fixed, stained with Alexa 546–phalloidin (red) and imaged for GFP (green) on a Zeiss
700 confocal microscope. Projection images were generated from stacks of 20 to 40 optical sections
collected at 1.2-mm intervals. Scale bar, 50 mm. (B) Viral transduction rates were determined from the
number of eGFP-positive hair cells (green) in each cochlea divided by the total hair cells with Alexa
546–phalloidin–positive hair bundles. Data are means ± SD (n, number of cochleas). Symbols show trans-
duction rates for each cochlea. (C) Viral transduction rates for all hair cells subdivided into five equal re-
gions and plotted for the entire length of the tonotopic axis. Data are means ± SD [n as shown in (B)]. (D)
Representative images of cochleas dissected from P0 WT mice, exposed to AAV2/1-eGFP vectors with
promoters indicated (titers: 1 × 1011
to 1 × 1012
gc/ml). Scale bars, 50 mm.

characterization of AAV2/1 transduction in vivo using eGFP as a
reporter.
Round window injection of AAV2/1 targets hair cells
To investigate expression of AAV vectors in the cochlea in vivo, we
developed a method for viral delivery to the perilymphatic spaces via
round window membrane (RWM) injection into early postnatal mice
(P0 to P2). Our RWM injection protocol is similar to other methods
(5), except that we left the bulla intact and inserted the micropipette
directly through the overlying fascia until it penetrated both the bulla
and RWM into the scala tympani. In initial experiments, successful
targeting of the perilymphatic spaces was confirmed, where dye-filled
turns of the entire cochlea were visually discernible through the bulla.
In the next series of experiments, we injected either AAV2/1-Cba-
eGFP or AAV2/1-Cmv-eGFP unilaterally into the left ear through the
RWM of P0 to P2 wild-type mice. When injected ears were harvested
at P8 to P10, eGFP fluorescence revealed that both AAV2/1-Cba-
eGFP (Fig. 2, A and B) and AAV2/1-Cmv-eGFP (fig. S1) vectors tar-
geted hair cells and supporting cells and drove transgene expression
throughout the cochlea. However, unlike the in vitro results, eGFP was
mainly expressed in IHCs in vivo. GFP-positive OHCs were seen spo-
radically in the basal half of the cochlea, near the injection site, but
very few GFP-positive OHCs were found in the apical half of the co-
chlea (Fig. 2, A and B). In cochleas injected with AAV2/1-Cba-eGFP,
59 ± 2% (±SD; n = 2) of IHCs were eGFP-positive, and with AAV2/
1-Cmv-eGFP, 70 ± 9% were eGFP-positive (n = 4). The utricles of in-
jected mice also displayed eGFP expression in vestibular hair cells and
supporting cells (fig. S2), which confirmed that the injections distributed
viral particles throughout the membranous labyrinth. We did not see
evidence of GFP expression upon gross inspection of the auditory
brainstem or other bodily tissues.
Both Cmv and Cba drove robust expression of eGFP in hair cells,
but we opted to focus on the Cba promoter, which has recently been
shown effective for driving exogenous Vglut3 expression in mouse
IHCs in vivo (5). To assay for possible deleterious consequences on
hair cell function, we measured sensory transduction currents from
AAV2/1-Cba-eGFP–transduced wild-type hair cells after RWM injec-
tions at P0 to P2. Whole-cell, tight-seal recording revealed sensory
transduction currents from eGFP-positive IHCs that were similar to
those of GFP-negative control hair cells from the same tissue (Fig. 2, C
to F) and similar to currents of uninjected control cells (11). The sen-
sitivity of the cells, as indicated by the steepness of the stimulus-response
relationship, was unaltered in the eGFP-positive cochlear hair cells rela-
tive to control cells (Fig. 2, D to F). Current amplitudes (Fig. 2E) and
adaptation properties (Fig. 2C) were also unaffected, suggesting that viral
transduction and eGFP expression do not alter sensory transduction.
Because sensory transduction has not been previously recorded after
the in vivo injection of viral vectors, these data offer assurance that
hair cell function is not compromised by AAV2/1 injection or eGFP
expression.
To assay for possible consequences of intracochlear injection of
AAV2/1-Cba-eGFP on auditory function, we recorded auditory brain-
stem responses (ABRs) from ears of injected and uninjected wild-type
mice at P25 (Fig. 2G). The ABR assay uses scalp electrodes to monitor
the summed electrical activity of the auditory brainstem, with the first
peak representing activity in the 8th cranial nerve. Consistent with
previous studies that showed no detrimental effect on ABRs after in
utero injection (18) or adult injection (19), auditory thresholds were
not significantly different between uninjected wild-type control mice,
wild-type mice that received a sham RWM injection with phosphate-
buffered saline (P = 0.27, t test), or wild-type mice that received RWM
injection that contained AAV2/1-Cmv-eGFP (P = 0.95, t test) (Fig. 2H).
In summary, neither the injection technique, AAV transduction, nor
eGFP expression affected hair cell or auditory function in any of our
assays, suggesting that AAV2/1 vectors are safe for delivery of exogenous
genes into the inner ears of neonatal mice.
AAV2/1-Cba-Tmc vectors rescue hair cell function in vitro
To assess the potential for gene therapy restoration of hair cell and
auditory function, we generated AAV2/1-Cba vectors that carried
the coding sequence for wild-type Tmc1 or Tmc2 fused to 3xFLAG
tags at their C termini. To evaluate the functionality of these vectors,
we applied the AAV2/1-Cba-Tmc1 or AAV2/1-Cba-Tmc2 vectors di-
rectly to organotypic cochlear cultures excised at P0 from Tmc1D/D
;
Tmc2D/D
mice, which lack TMC1 and TMC2 protein expression, are
deaf, and lack sensory transduction in both OHCs (10) and IHCs (11).
After 5 to 7 days, hair cells of the vector-exposed Tmc1D/D
;Tmc2D/D
cultures recovered FM1-43 uptake (fig. S3A), a styryl dye that perme-
ates transduction channels open at rest (20–22). Because uninfected
hair cells from Tmc1D/D
;Tmc2D/D
do not take up FM1-43 (10), dye
uptake in cells exposed to AAV2/1-Cba-Tmc vectors indicates recov-
ery of sensory transduction.
Sensory transduction currents were recorded from both IHCs and
OHCs. Figure S3B shows representative currents recorded from Tmc1D/D
;
Tmc2D/D
hair cells and hair cells in the same tissue 5 to 7 days after
exposure to AAV2/1-Cba-Tmc1. Peak sensory transduction currents
from OHCs ranged in amplitude from 66 to 420 pA, and those from
IHCs ranged from 50 to 800 pA (fig. S3B). The average (±SD) peak
transduction current for IHCs rescued by AAV2/1-Cba-Tmc1 was
306 ± 211 pA (n = 4), and that for OHCs, 289 ± 98 pA (n = 10).
For IHCs rescued by AAV2/1-Cba-Tmc2, the mean peak transduction
current was 766 ± 142 pA (n = 2). These results demonstrate that either
Tmc1 or Tmc2 can restore sensory transduction at the cellular level
when delivered into nonfunctional hair cells in vitro.
AAV2/1-Cba-Tmc vectors restore sensory transduction in vivo
To evaluate the ability of AAV2/1-Cba-Tmc1 vectors to drive exoge-
nous expression of Tmc1 in vivo, we used quantitative reverse transcrip-
tion polymerase chain reaction (RT-PCR) with primers specific for
Tmc1 mRNA (10). Tmc1D/D
mice were injected at P1 with AAV2/
1-Cba-Tmc1 at a titer of 2 × 1013
gc/ml into one ear. RNA was har-
vested from injected and uninjected cochleas at P14. Injected cochleas
had Tmc1 mRNA expression levels that were 12-fold higher than those
in uninjected cochleas (Fig. 3A), consistent with AAV2/1-Cba-Tmc1–
driven expression of exogenous Tmc1.
To test the ability of the AAV-Cba-Tmc vectors to drive TMC pro-
tein expression in hair cells in vivo, we injected Tmc1D/D
;Tmc2D/D
mice
at P0 to P2 and excised cochlear tissue 6 to 7 days later. We observed
prominent Tmc1-3xFLAG staining in the cell bodies of most IHCs
(Fig. 3, B and C) and at the tips of IHC stereocilia (Fig. 3D), the site
of hair cell sensory transduction, which confirmed that TMC1 and
TMC2 were expressed and properly targeted. Consistent with our pre-
vious in vivo observation (Fig. 2 and fig. S1), we saw little expression
of the exogenous protein in OHCs (Fig. 3, B to D).
To assay for rescue of FM1-43 uptake and hair cell sensory transduc-
tion, live cochlea were excised at P6 to P7 and maintained in organotypic

cultures until the equivalent of P8 to P9. FM1-43FX was pipetted over
the tissue. Figure 3E shows a representative image of a cochlear culture
harvested from a Tmc1D/D
;Tmc2D/D
mouse exposed to FM1-43FX, which
reveals no dye uptake in IHCs or OHCs. In contrast, Tmc1D/D
;Tmc2D/D
mice injected with AAV2/1-Cba-Tmc1 vectors had robust FM1-43FX
uptake in most IHCs (71%, 57 of 80 cells) along the entire length
of the cochlea in the injected ear (Fig. 3F). Very few OHCs took up
the dye, consistent with a lack of viral transduction in OHCs in vivo.
Because FLAG and FM1-43 labeling indicated high viral transduc-
tion rates in IHCs in AAV2/1-Cba-Tmc1–injected cochleas, we assayed
Fig. 2. In vivo injection of AAV2/1-Cba-eGFP through the RWM. (A) Rep-
resentative confocal images from the apex and base of a WT cochlea
injected through the RWM with 1 ml of AAV2/1-Cba-eGFP (6 × 1012
gc/ml)
at P2, harvested at P9, and stained with Alexa 546–phalloidin (red) and
imaged for GFP (green). Scale bar, 100 mm. (B) Apex and base from the
same cochlea in (A) at higher magnification. Scale bar, 50 mm. (C) Families
of sensory transduction currents evoked by mechanical displacement of
IHC bundles from control (GFP-negative) cells and GFP-positive cells. Scale
bars and displacement protocols are provided. (D) Stimulus-response curves
for GFP-negative and GFP-positive cells revealed no difference in sensitivity.
(E and F) Peak sensory transduction currents (E) and 10 to 90% operating
range (F) from control and GFP-positive cells. Data are means ± SD (n, num-
ber of cells). (G) Families of ABR waveforms recorded at P25 from uninjected
WT and AAV2/1-Cba-eGFP injected ears. The stimulus was an 8-kHz tone
burst between 25 and 70 dB in 5-dB increments. (H) Auditory thresholds
plotted as a function of stimulus frequency for uninjected WT mice, sham-
injected WT mice, and AAV2/1-Cba-eGFP–injected mice. Data are means ±
SD (n, number of mice).

for rescue of sensory transduction current in cochlear IHCs at P7 to
P9. Sensory transduction currents were recorded from AAV-Cba-
Tmc–positive IHCs that expressed exogenous Tmc1 or Tmc2 (Fig. 3G),
identified by the presence of FM1-43 uptake. Although the FM1-43
uptake data revealed variable viral transduction throughout the co-
chlea, cells that were FM1-43–positive had normal sensory transduc-
tion current amplitudes (Fig. 3, G and H) and normal sensitivity (Fig. 3,
I and J), relative to wild-type (11) and GFP-positive controls (Fig. 2, C
to F). FM1-43–negative cells from the same ear lacked sensory trans-
duction currents entirely (Fig. 3, G and H). The differences in adapta-
tion rate and extent between hair cells exposed to AAV2/1-Cba-Tmc1
and AAV2/1-Cba-Tmc2 (Fig. 3G) were consistent with the differences
observed in hair cells expressing endogenous Tmc1 or Tmc2 (11). In sum-
mary, the single hair cell physiology data suggest that AAV2/1-Cba-Tmc
vectors are capable of complete restoration of sensory transduction in
vivo with all the properties of native sensory transduction.
AAV2/1-Cba-Tmc1 rescues ABRs in Tmc1-deficient mice
To model AAV gene therapy for rescue of genetic deafness in pa-
tients who carry recessive mutations in Tmc1, we injected Tmc1-deficient
animals (P0 to P2) in vivo with AAV2/1-Cba-Tmc1 and measured the
auditory function at 25 to 30 days. Figure 4A shows families of ABR
waveforms recorded in response to 8-kHz tone bursts. The data were
recorded from an uninjected Tmc1D/D
mouse and a Tmc1D/D
mouse
injected with AAV2/1-Cba-Tmc1. Uninjected Tmc1D/D
mice lacked re-
sponses at all stimulus intensities and frequencies tested, which ranged
between 0 and 115 dB and 5 and 32 kHz, respectively, indicating pro-
found deafness, consistent with previous reports (9). However, prom-
inent ABR waveforms, which represented substantial recovery of auditory
transmission from the cochlea to the brainstem via the 8th cranial
Fig. 3. Exogenous, AAV-delivered Tmc1/2 restores sensory transduc-
tion in Tmc-deficient hair cells in vivo. P0 to P2 Tmc1D/D
;Tmc2D/D
mice
were injected via the RWM with AAV2/1-Cba-Tmc1 (2.4 × 1013
gc/ml) or
AAV2/1-Cba-Tmc2 (1.8 × 1013
gc/ml). Cochleas were harvested 6 to 7 days
after injection. (A) Quantitative RT-PCR expression analysis of Tmc1 mRNA
from total RNA harvested from two uninjected Tmc1D/D
cochleas and two
Tmc1D/D
cochleas injected with AAV2/1-Cba-Tmc1 (n = 3 technical replicates).
(B) Percent TMC1-FLAG–positive hair cells in AAV2/1-Cba-Tmc1–injected
cochleas (n, number of FLAG-positive cells over total number of cells).
(C) Confocal image of a cochlea injected with AAV2/1-Cba-Tmc1 and stained
with Alexa 488 anti-FLAG antibody (green) and Alexa 546–phalloidin (red).
Scale bar, 50 mm. (D) Projection from z-stack images of a WT cochlea in-
jected with AAV6/1-Cba-Tmc2 showing FLAG staining at the tips of hair cell
stereocilia. Scale bar, 5 mm. (E and F) FM1-43 uptake in Tmc1D/D
;Tmc2D/D
tissue not exposed to AAV2/1-Cba-Tmc vectors (control; E) or injected (F)
with AAV2/1-Cba-Tmc1. OC, organ of Corti. Scale bar, 50 mm. (G) Represent-
ative families of sensory transduction currents recorded from IHCs of a
Tmc1D/D
;Tmc2D/D
mouse injected with AAV2/1-Cba-Tmc1 that were FM1-
43–negative (left) or FM1-43–positive (middle). FM1-43–positive IHC currents
from a Tmc1D/D
;Tmc2D/D
mouse injected with AAV2/1-Cba-Tmc2 (right). (H)
Peak sensory transduction current amplitudes from FM1-43–negative and
FM1-43–positive IHCs of Tmc1D/D
;Tmc2D/D
mice injected with AAV2/1-Cba-
Tmc1 or AAV2/1-Cba-Tmc2 as indicated. Bars are means ± SD. Circles are in-
dividual measurements (n, number of cells). (I) Stimulus-response curves from
the currents shown in (G). (J) Ten to 90% operating range measured from
stimulus-response curves in (I). Bars are means ± SD. Circles are individual mea-
surements (n, number of cells).

nerve, were present in 50% (8 of 16) of the AAV2/1-Cba-Tmc1–
injected mice. In the eight mice with no recovery of ABR responses,
we found little evidence of viral transduction in hair cells, suggesting
that the injections may have been unsuccessful, perhaps due to
clogged or improperly targeted injection needles.
For the eight AAV2/1-Cba-Tmc1–injected mice with auditory re-
sponses, we quantified peak 1 amplitudes as a function of stimulus in-
tensity at 8 kHz and compared them with those observed in uninjected
Tmc1D/D
control mice (Fig. 4B). Peak 1 amplitudes increased monoton-
ically in the eight Tmc1D/D
mice injected with AAV-Cba-Tmc1, indicat-
ing a stimulus-dependent increase in the auditory response. Minimum
ABR thresholds showed recovery of auditory function in these mice,
particularly at frequencies between 5 and 16 kHz (Fig. 4C). The ABR
thresholds at 85 to 100 dB (Fig. 4C) represent a substantial improve-
ment relative to uninjected Tmc1D/D
control mice, which are profoundly
deaf and have no detectable responses to sound stimuli even at 115 dB,
the highest intensity tested. The data demonstrate partial recovery of
auditory function at the systems level in the otherwise deaf mice.
Although there was recovery of ABR responses in the AAV2/
1-Cba-Tmc1–injected mice, the responses did not reach wild-type
levels (Fig. 2H). To investigate the reason for the incomplete recovery,
including possible toxicity associated with overexpression of Tmc1, we
injected wild-type C57BL/6 mice with AAV2/1-Cba-Tmc1. ABR thresh-
olds in the AAV2/1-Cba-Tmc1–injected wild-type mice were unaltered
relative to uninjected controls (fig. S4A). This finding suggests that there
is little toxicity associated with the injection procedure, exposure to
AAV2/1-Cba-Tmc1 vectors, or overexpression of Tmc1 in hair cells,
spiral ganglion neurons, or any other cell type necessary for normal
auditory function. Furthermore, we did not observe FM1-43 uptake
in non-hair cells (Fig. 3, E and F), suggesting that aberrant expres-
sion of Tmc1 does not lead to the formation of functional channels in
other cell types.
To investigate other possible causes of the incomplete recovery, we
measured distortion product otoacoustic emissions (DPOAEs) in un-
injected Tmc1D/D
mice and Tmc1D/D
mice injected with AAV2/1-Cba-
Tmc1. The DPOAE assay is a specific test for OHC function. OHCs are
required for cochlear amplification, enhanced sensitivity, and normal
auditory function (23, 24). DPOAE measurements revealed elevated
thresholds relative to wild-type mice and no difference between un-
injected Tmc1D/D
mice and those injected with AAV2/1-Cba-Tmc1
vectors (Fig. 4D), which suggests little recovery of OHC function. Un-
injected wild-type mice and wild-type mice injected with either AAV2/
1-Cba-eGFP or AAV2/1-Cba-Tmc1 had normal DPOAE thresholds
(fig. S4B), consistent with the suggestion that the injection itself, ex-
posure to AAV2/1 vectors, and Tmc1 overexpression caused little
toxicity in OHCs or elsewhere in the cochlea.
After the ABR and DPOAE measurements, the mice were eutha-
nized and their inner ear tissue was excised for histological examina-
tion. There was no overt evidence of inflammation, tissue damage, or
Fig. 4. Exogenous Tmc1 rescues auditory function in Tmc1D/D
mice. (A)
Families of ABR waveforms recorded from an uninjected Tmc1D/D
mouse
and from a Tmc1D/D
mouse injected with AAV2/1-Cba-Tmc1. ABRs were re-
corded at P25 to P30 using 8-kHz tone bursts at sound pressure levels be-
tween 75 and 105 dB in 5-dB increments. Scale bar applies to both families.
(B) Peak 1 amplitudes measured from 8-kHz ABR waveforms, as shown in
(A), for eight Tmc1D/D
mice injected with AAV2/1-Cba-Tmc1 vectors. Open
circles are mean responses (± SD) from uninjected Tmc1D/D
mice (n = 8). (C)
ABR thresholds plotted as a function of sound frequency for eight Tmc1D/D
mice injected with AAV2/1-Cba-Tmc1 vectors. Open circles are means of
uninjected Tmc1D/D
mice at the highest sound intensity tested (arrows)
(n = 8). (D) DPOAE thresholds as a function of stimulus frequency for WT,
uninjected Tmc1D/D
mice and Tmc1D/D
mice injected with AAV2/1-Cba-Tmc1.
Data are means ± SD (n, number of animals). (E) Percentage of surviving
IHCs (relative to WT) in 5-mm mid-cochlea sections from Tmc1D/D
mice and
AAV2/1-Cba-Tmc1–injected mice (upper n, number of IHCs; lower n, num-
ber of cochlea). (F) Confocal images of cochlear whole mounts harvested at
P30 from an uninjected Tmc1D/D
mouse and a Tmc1D/D
mouse injected with
AAV2/1-Cba-Tmc1. The tissue was stained for MYO7A (green) and phalloidin
(red). Scale bar, 50 mm. Figure S5 shows low-magnification images of the
same cochleas.

decay in the injected ears. Cochlear whole mounts were stained with
an anti-MYO7A antibody to label hair cells and with Alexa546–
phalloidin to label hair bundles (Fig. 4F). Counts of surviving IHCs re-
vealed no significant difference (P = 0.6) between uninjected Tmc1D/D
cochleas and those injected with AAV2/1-Cba-Tmc1 (Fig. 4E), suggest-
ing that AAV2/1-Cba-Tmc1 injection was neither detrimental nor ben-
eficial for IHC survival at P30. Hair bundle morphology and tip links
remain normal in surviving Tmc1D/D
hair cells (10), and we detected no
changes after AAV2/1-Cba-Tmc1 injection.
AAV2/1-Cba-Tmc2 restores ABRs in Tmc1 mutant mice
Tmc2, a closely related ortholog of Tmc1, is expressed in cochlear hair
cells during the first postnatal week, but mRNA levels decline there-
after (10). Tmc2 is expressed in Tmc1D/D
mice, and both IHCs and
OHCs in these mice retain sensory transduction through the end of
the first postnatal week (10, 11). However, because Tmc2 expression
declines thereafter, Tmc1D/D
mice are deaf. AAV2/1-Cba-Tmc2 restored
sensory transduction in Tmc1D/D
;Tmc2D/D
hair cells in vitro and in vivo
(Fig. 3, G and H); therefore, we investigated whether AAV2/1-Cba-
Tmc2 injection into the ears of Tmc1D/D
mice would also restore auditory
function. To test this possibility, we used the same injection protocols (as
AAV2/1-Cba-Tmc1) and measured ABR responses at P25 to P30. We
recorded prominent ABR responses in 9 of 16 Tmc1D/D
mice injected
with AAV2/1-Cba-Tmc2 (Fig. 5A). Peak 1 amplitudes (Fig. 5B) and
minimal ABR thresholds (Fig. 5C) were similar to those of mice injected
with AAV2/1-Cba-Tmc1 (Fig. 4, B and C), suggesting that expression of
exogenous Tmc2 is capable of restoring auditory function in vivo. Also
similar to AAV2/1-Cba-Tmc1, AAV2/1-Cba-Tmc2 injection did not
restore DPOAE thresholds (fig. S4B), nor did it affect hair cell survival
or death rates relative to uninjected Tmc1D/D
control mice (Fig. 5D
and fig. S5). There was no significant difference (P = 0.44) in IHC counts
between uninjected Tmc1D/D
mice and those injected with AAV2/
1-Cba-Tmc2 (Fig. 5E).
On the basis of the ability of exogenous Tmc2 to restore ABR re-
sponses in mouse models of genetic deafness due to recessive TMC1
mutations, we next asked whether exogenous Tmc2 expression might
be sufficient to overcome dominant TMC1 mutations and restore au-
ditory function. Previously, Pan et al. (11) showed that the dominant
p.M412K mutation in TMC1, known as Bth, causes a reduction in
calcium permeability, a reduction in single-channel currents, and an
increase in the number of sensory transduction channels in IHCs. The
identical mutation has also recently been described in a human family
in the orthologous residue of human TMC1 p.M418K (15), suggesting
that the Bth mouse is an ideal model for genetic hearing loss in
humans. To investigate the ability of Tmc2 to compensate for the
Tmc1-Bth mutation, we injected AAV2/1-Cba-Tmc2 vectors into the
ears of homozygous Bth mice and measured their ABRs at P25 to P30.
Consistent with previous reports, Bth mice were completely deaf by
P15 (13, 14). However, 7 of 15 Bth mice injected with AAV2/1-Cba-
Tmc2 had prominent ABRs evoked by loud sound intensities (Fig. 6A).
Peak 1 amplitudes were smaller and the thresholds were elevated (Fig. 6,
B and C) relative to ABRs evoked in Tmc1D/D
mice injected with either
AAV2/1-Cba-Tmc2 (Fig. 5, B and C) or AAV2/1-Cba-Tmc1 (Fig. 4, B
and C), suggesting that the extent of the recovery was limited. There
was a significant increase in the survival rate of IHCs in the ears of
Bth mice injected with AAV2/1-Cba-Tmc2 relative to the uninjected
ears (Fig. 6, D and E, and fig. S6), suggesting that exogenous expres-
sion of Tmc2 promotes IHC survival.
AAV2/1-Cba-Tmc vectors restore behavioral responses to
auditory stimuli
Partial ABR recovery in deaf mice injected with AAV2/1-Cba-Tmc
vectors may have resulted in behaviorally relevant sound perception.
We therefore tested acoustic startle reflexes at P30 from wild-type
control mice, uninjected deaf mice, and deaf mice injected with
Fig. 5. Exogenous Tmc2 rescues auditory function in Tmc1D/D
mice. (A)
Families of ABR waveforms recorded from a Tmc1D/D
mouse injected with
AAV2/1-Cba-Tmc2. ABRs were recorded at P25 to P30 using 8-kHz tone
bursts at sound pressure levels between 75 and 105 dB in 5-dB increments.
Scale bar applies to both families. (B) Peak 1 amplitudes measured from 8-kHz
ABR waveforms, as shown in (A), for six Tmc1D/D
mice injected with AAV2/
1-Cba-Tmc2 vectors. Open circles are mean responses (±SD) from unin-
jected Tmc1D/D
mice (n = 8). (C) ABR thresholds plotted as a function of sound
frequency for six Tmc1D/D
mice injected with AAV2/1-Cba-Tmc2 vectors.
Open circles are means of uninjected Tmc1D/D
mice at the highest sound
intensity tested (arrows) (n = 8). (D) Confocal images of cochlear whole
mounts harvested at P30 from a Tmc1D/D
mouse injected with AAV2/
1-Cba-Tmc2. The tissue was stained for MYO7A (green) and phalloidin
(red). Scale bar, 50 mm. Figure S5 shows low-magnification images of the
same cochleas. (E) Percentage of surviving IHCs in 5-mm mid-cochlea
sections from Tmc1D/D
mice and AAV2/1-Cba-Tmc2–injected mice (upper
n, number of IHCs; lower n, number of cochleas).

AAV2/1-Cba-Tmc vectors. Wild-type mice had startle responses that
were detectable beginning around 80 dB (Fig. 7A), whereas Tmc1D/D
deaf mice lacked startle responses at all frequencies and sound inten-
sities tested (Fig. 7B). When we measured acoustic startle responses in
five Tmc1D/D
mice injected with AAV2/1-Cba-Tmc1 that lacked ABR
responses, we found that they also lacked startle responses (Fig. 7A,
open squares), perhaps due to failed injections. However, we found
that 100% (n = 7) of Tmc1D/D
mice injected with AAV2/1-Cba-Tmc1
that had ABR responses also had robust startle responses (Fig. 7A). All
seven mice had recovery of startle response thresholds at 90 to 100 dB
with startle amplitudes that increased with increasing sound intensity.
Furthermore, we found that the startle responses were also present at
P60, the latest time point tested: five of five Tmc1D/D
mice injected
with AAV2/1-Cba-Tmc1 that had positive ABRs also had substantial
startle responses (Fig. 7B).
Next, we examined startle responses in Tmc1 mutant mice injected
with AAV2/1-Cba-Tmc2. Although five Tmc1D/D
mice injected with
AAV2/1-Cba-Tmc2 recovered ABRs and were tested for startle re-
sponses, only two of the five had startle responses (Fig. 7C). One
mouse had substantial responses, similar to those of wild-type mice
for the loudest sound intensities (Fig. 7C). Seven of 15 Tmc1Bth
mice
injected with AAV2/1-Cba-Tmc2 recovered ABR responses, but none
had startle responses (Fig. 7, C and D), suggesting that exogenous
Tmc2 expression may not be sufficient to overcome the dominant
Bth mutation in a behaviorally relevant assay.
Figure 7D summarizes the ABR and startle response data for all 55
Tmc1 mutant mice injected with AAV2/1-Cba-Tmc vectors. Gene
augmentation with wild-type Tmc1 successfully restored both ABR and
acoustic startle responses in more than 50% of the injected Tmc1D/D
mice.
Injection of the AAV2/1 vector encoding wild-type Tmc2 revealed similar
success rates for ABR recovery, but was less effective for the recovery
of startle responses in both Tmc1D/D
and Tmc1Bth
mice.
DISCUSSION
To model gene therapy for DFNB7/11, we used mice deficient in
Tmc1 and engineered AAV2/1 vectors that carried the Cba promoter
and the coding sequence for either Tmc1 or Tmc2. Viral transduction
with either AAV2/1-Cba-Tmc1 or AAV2/1-Cba-Tmc2 revealed local-
ization of exogenous TMC1 and TMC2 at the tips of hair cell stereo-
cilia, uptake of the transduction channel permeable dye FM1-43, and
robust mechanosensory transduction currents in otherwise nonfunc-
tional hair cells. The data provide compelling evidence that the vectors
can restore function at the cellular level in vitro.
In vivo injection, via the RWM, of either AAV2/1-Cmv-eGFP or
AAV2/1-Cba-eGFP drove robust expression of eGFP in cochlear and
vestibular hair cells, suggesting that the approach may be viable for the
delivery of therapeutic reagents to target hair cells throughout the hu-
man inner ear. Normal mechanosensory responses and normal ABRs
suggested that the injection technique, AAV2/1 vectors, and eGFP ex-
pression are safe for in vivo use, supporting further development of
AAV gene therapy as a strategy for hearing restoration. Consistent with
previous observations (5), when AAV2/1 vectors were injected via the
RWM into the inner ears of deaf mice, restoration of cellular function
was limited to IHCs. We found little evidence of exogenous gene ex-
pression in OHCs after injection of four different vectors. Because all
vectors were capable of driving exogenous gene expression in OHCs
Fig. 6. Exogenous Tmc2 rescues au-
ditory function in Tmc1-Bth mice. (A)
Families of ABR waveforms recorded
from a Tmc1-Bth mouse and a Tmc1-Bth
mouse injected with AAV2/1-Cba-Tmc2.
ABRs were recorded at P25 to P30 using
8-kHz tone bursts at sound pressure
levels between 80 and 110 dB in 5-dB in-
crements. Scale bar applies to both fam-
ilies. (B) Peak 1 amplitudes measured from
8 kHz ABR waveforms, as shown in (A),
for seven Tmc1-Bth mice injected with
AAV2/1-Cba-Tmc2 vectors. Open circles are mean responses (±SD) from un-
injected Tmc1-Bth mice (n = 5). (C) ABR thresholds plotted as a function of
sound frequency for seven Tmc1-Bth mice injected with AAV2/1-Cba-Tmc2
vectors. Open circles are means of uninjected Tmc1-Bth mice at the highest
sound intensity tested (arrows) (n = 5). (D) Confocal images of cochlear whole
mounts harvested at P30 from an uninjected Tmc1-Bth mouse and a Tmc1-
Bth mouse injected with AAV2/1-Cba-Tmc2. The tissue was stained for
MYO7A (green) and phalloidin (red). Scale bar, 100 mm. Note the increased
survival of IHCs in the apex and base of the AAV2/1-Cba-Tmc2 injected
cochlea. Figure S6 shows low-magnification images of the same cochleas.
(E) Percentage of surviving IHCs in Tmc1-Bth mice and AAV2/1-Cba-Tmc2–
injected Bth mice (upper n, number of IHCs; lower n, number of cochleas).

in vitro, we suspect that the lack of viral transduction in OHCs in vivo
resulted from limited viral access to the hair cell apical surface via
RWM injection into perilymphatic spaces. Kilpatrick et al. (19) re-
ported that introduction of AAV vectors into the scala media, which
bathes hair cell apical membranes, yielded GFP expression in both
IHCs and OHCs. The challenge of scala media injection is that it re-
quires a more invasive surgical approach and can cause mixing of high
K+
(~140 mM) endolymph and perilymph leading to hair cell de-
polarization and cell death. To target OHCs may require vectors that
can enter via the basolateral membrane or delivery methods that target
endolymphatic spaces without disrupting endolymph/perilymph
barriers.
ABRs were recovered in >50% of Tmc1D/D
deaf mice injected with
AAV2/1-Cba-Tmc1, indicating successful transmission of auditory
information from the cochlea to the au-
ditory brainstem. The ABR thresholds
were elevated relative to those of wild-
type mice, indicating incomplete recovery
of auditory function. DPOAE responses
did not recover, suggesting that the elevated
ABR thresholds were due to lack of re-
covery of OHC function, in turn due to
low viral transduction rates in OHCs.
Functional OHCs are required for cochlear
amplification, a process that provides me-
chanical feedback to the cochlea by increas-
ing gain to soft sounds. OHC dysfunction
is known to yield elevated ABR thresh-
olds, shifted up to 60 dB higher than
wild type. Thus, in Tmc1D/D
mice, in which
all cochlear hair cells lack sensory trans-
duction, rescue of IHC but not OHC func-
tion yielded ABR thresholds ~60 dB higher
than wild type, similar to thresholds in
mice with OHC dysfunction (24). In Vglut3
knockout mice, OHCs remain functional
but the mice are deaf because of IHC dys-
function (5). After Vglut3 gene augmen-
tation, ABR thresholds recovered to near
wild-type levels because restoration of
function was only required in IHCs, which
account for ~25% of the cochlear hair cell
population. Our experiments revealed
high viral transduction rates in IHCs
and the transduced cells had mechanosen-
sory currents equivalent to those of wild
type, but the OHC dysfunction remained.
Although the recovery was incomplete, the
result was considered a success because
normal mechanosensory function in IHCs
is a prerequisite for auditory function. Had
the outcome been the converse—restoration
of OHC but not IHC function—the animals
would still be deaf.
We also found that AAV2/1-Cba-Tmc2
vectors were capable of restoring sensory
transduction and partial ABR responses in
Tmc1D/D
mice, which supports the hypoth-
esis that Tmc1 and Tmc2 perform somewhat redundant functions and
can substitute for each other, at least in IHCs. The AAV2/1-Cba-Tmc2
transduction pattern was similar to AAV2/1-Cba-Tmc1 and was re-
stricted primarily to IHCs, resulting in similar recovery at elevated ABR
thresholds. That hair cell survival rates were not altered in Tmc1D/D
mice
injected with either AAV2/1-Cba-Tmc1 or AAV2/1-Cba-Tmc2 was im-
portant for two reasons: (i) neither vector caused loss or decay of hair cells,
and (ii) hair cells remained in uninjected Tmc1D/D
mice up to P60, suggest-
ing that there may be a window of opportunity for therapeutic interven-
tion. Whether a similar window exists in humans with recessive TMC1
mutations is unknown. If patients with TMC1 mutations retain viable
hair cells, they may present an opportunity for clinical intervention.
Restoration of auditory function was limited in Bth mice injected
with AAV2/1-Cba-Tmc2. There was significant preservation of IHCs
Fig. 7. Exogenous Tmc expression rescues acoustic startle responses in Tmc1 mutant mice. (A) Star-
tle response amplitudes measured at P30 and plotted as a function of sound intensity and as mean ± SD
of four control C57BL/6 mice (open circles), seven individual Tmc1D/D
mice injected with AAV2/1-Cba-Tmc1,
and five AAV2/1-Cba-Tmc1–injected mice with no recovery (open squares). (B) Startle responses measured at
P60 and plotted as mean ± SD of four Tmc1D/D
mice (open circles) and five individual Tmc1D/D
mice injected
with AAV2/1-Cba-Tmc1. (C) Startle responses measured at P30, plotted for two individual Tmc1D/D
mice
injected with AAV2/1-Cba-Tmc2 and as mean ± SD of seven Tmc1-Bth mice (open circles) injected with
AAV2/1-Cba-Tmc2. (D) Summary bar graph showing the percentage of Tmc1 mutant mice with recovery
as assayed by ABRs and startle responses for mice injected with either AAV2/1-Cba-Tmc1 or AAV2/1-Cba-Tmc2.
Numerator indicates n mice with recovery of function; denominator indicates n injected mice tested. Not
all mice were tested with both assays.

in AAV2/1-Cba-Tmc2–injected Bth mice. The mechanism that pro-
moted IHC survival is unknown. On the basis of the measurements
of sensory transduction and calcium permeability in mice that ex-
pressed wild-type Tmc2, Tmc1, or Tmc1-Bth, Pan et al. (11) found a
significant reduction in calcium entry in Tmc1-Bth IHCs, whereas
Tmc2 cells had high calcium entry. We hypothesize that appropriate
levels of calcium entry are required for maintenance and survival of
IHCs. Therefore, by introducing exogenous Tmc2, calcium homeostasis
was restored, which enhanced hair cell survival in the Bth mice injected
with AAV2/1-Cba-Tmc2.
As a final test of auditory function, we measured acoustic startle
reflexes in Tmc1 mutant mice. The otherwise unresponsive Tmc1D/D
mice recovered startle responses after injection of AAV2/1-Cba-Tmc1,
and the responses persisted for up to 60 days, the latest time point
tested. It was unclear why Tmc1-Bth mice injected with AAV2/1-Cba-
Tmc2 recovered partial ABR function but did not recover startle re-
sponses. The extent of the ABR recovery in Tmc1-Bth mice injected
with AAV2/1-Cba-Tmc2 was less than the ABR recovery in Tmc1D/D
mice injected with AAV2/1-Cba-Tmc2, suggesting that there may be a
minimal threshold required to drive behavioral responses to loud
sounds. Therapies aimed at restoration of auditory function for dom-
inant DFNA36 deafness may require development of alternate strate-
gies, perhaps by suppression of the dominant allele.
In conclusion, the data provide compelling proof-of-principle evi-
dence demonstrating that gene augmentation in a mouse model of
DFNB7/11 is effective in restoring cellular function in vitro in both
IHCs and OHCs, restoring IHC function in vivo, partial recovery of
systems level function in vivo, and recovery of acoustic startle reflexes
at the behavioral level. Recovery of ABR and startle responses was
likely a direct result of recovery of IHC sensory transduction at the
cellular level and suggests that Tmc1 reexpression can restore auditory
function at every level.
Thirty-five TMC1 mutations have been identified that cause re-
cessive prelingual deafness in humans, which underscores the signifi-
cance of TMC1 for normal auditory function and the need for
therapeutic reagents to remedy the disorder. Although our gene ther-
apy strategy is not yet ready for clinical application, the challenges that
remain are not insurmountable. Continued development of Tmc gene
therapy will need to provide characterization of the long-term expres-
sion pattern of the exogenous constructs, including their ability to
maintain recovery; improved design of vectors, promoters, and delivery
techniques that drive exogenous gene expression in OHCs; and further
evaluation of the therapeutic window of opportunity in humans with
recessive TMC1 mutations. Finally, we suggest that AAV-mediated gene
augmentation in the inner ear may be a model that could be expanded
to address some of the more than 70 forms of genetic deafness.
MATERIALS AND METHODS
Study design
The aim of this study was to identify AAV serotypes and promoters
for delivery and expression of exogenous Tmc1 and Tmc2 in hair cells
of the mouse cochlea and to evaluate the ability of these vectors to
restore function in mouse models of genetic deafness in humans. AAV
vectors were injected in vivo, and the outcomes were evaluated using
quantitative RT-PCR, immunolocalization and confocal microscopy,
imaging FM1-43 uptake, single-cell recording, histology and imaging
of whole cochleas, measurement of ABRs, DPOAEs, and acoustic startle
reflexes. Left ears were injected and right ears were used as uninjected
controls. Each experiment was replicated as indicated by n values in the
figure legends. All experiments with mice and viral vectors were ap-
proved by the Institutional Animal Care and Use Committee (protocols
#2146 and #2878) at Boston Children’s Hospital and the Institutional
Biosafety Committee (protocol #IBC-P00000447).
In vivo injection of viral vectors
Mouse pups (P0 to P2) were injected via the RWM using beveled
glass microinjection pipettes, as described in Supplementary
Methods.
Hair cell electrophysiology
Organotypic cochlear cultures were bathed in standard artificial peri-
lymph containing 137 mM NaCl, 0.7 mM NaH2PO4, 5.8 mM KCl,
1.3 mM CaCl2, 0.9 mM MgCl2, 10 mM Hepes, and 5.6 mM D-glucose.
Vitamins (1:50) and amino acids (1:100) were added to the solution
from concentrates (Invitrogen), and NaOH was used to adjust the fi-
nal pH to 7.40 (310 mosmol/kg). Recording pipettes (3 to 5 megohms)
were pulled from R6 capillary glass (King Precision Glass) and filled
with intracellular solution containing 135 mM CsCl, 5 mM Hepes,
5 mM EGTA, 2.5 mM MgCl2, 2.5 mM Na2–adenosine triphosphate,
and 0.1 mM CaCl2, where CsOH was used to adjust the final pH to
7.40 (285 mosmol/kg). Whole-cell, tight-seal voltage-clamp recordings
were done at −84 mV at room temperature (22° to 24°C) using an
Axopatch 200B amplifier (Molecular Devices). Sensory transduction
currents were filtered at 10 kHz with a low-pass Bessel filter and digit-
ized at ≥20 kHz with a 16-bit acquisition board (Digidata 1440A) and
pCLAMP 10 software (Molecular Devices). Data were stored for off-
line analysis using OriginPro 8 (OriginLab).
ABR and DPOAE
ABR recordings were conducted as described previously (25), at 32°C
in a soundproof chamber. To test hearing function, anesthetized mice
were presented pure tone stimuli of 5.6, 8, 11.3, 16, 22.6, or 32 kHz at
sound pressure levels between 10 and 115 dB in 5-dB steps until a
threshold intensity that evoked a reproducible ABR waveform (peaks
1 to 4) was detected. Responses were collected, and data were analyzed
as described in Supplementary Methods.
DPOAE data were collected under the same conditions and during
the same recording sessions as the ABR data. Primary tones were
produced at a frequency ratio of 1.2 (f2/f1) for the generation of DPOAEs
at 2f1–f2, where the f2 level was 10 dB sound pressure level below f1
level for each f2/f1 pair. The f2 levels were swept in 5-dB steps from 20 to
80 dB. Waveform and spectral analyses are described in Supplemen-
tary Methods.
Acoustic startle reflexes
Mice were tested for startle reflexes in response to broadband auditory
stimulation at varying intensities, as described in Supplementary
methods.
Statistical analysis
All mean values and error bars presented in the figures represent
mean ± SD. Comparisons for statistical significance between injected
ears and uninjected ears were performed using a two-tailed paired t test.
P < 0.05 was considered significant.

SUPPLEMENTARY MATERIALS
www.sciencetranslationalmedicine.org/cgi/content/full/7/295/295ra108/DC1
Materials and Methods
Fig. S1. In vivo injection of AAV2/1-Cmv-eGFP through the RWM.
Fig. S2. In vivo injection also targets vestibular hair cells.
Fig. S3. Exogenous Tmc1/2 restores sensory transduction in Tmc-deficient hair cells in vitro.
Fig. S4. Auditory function in wild-type mice injected with AAV vectors.
Fig. S5. Confocal images of Tmc1D/D
cochleas injected with AAV2/1-Cba-Tmc vectors.
Fig. S6. Confocal images of Tmc1-Bth cochleas injected with AAV2/1-Cba-Tmc2.
References (26, 27)
REFERENCES AND NOTES
1. G. S. G. Géléoc, J. R. Holt, Sound strategies for hearing restoration. Science 344, 1241062 (2014).
2. D. C. Kohrman, Y. Raphael, Gene therapy for deafness. Gene Ther. 20, 1119–1123 (2013).
3. B. W. Kesser, G. T. Hashisaki, K. Fletcher, H. Eppard, J. R. Holt, An in vitro model system to
study gene therapy in the human inner ear. Gene Ther. 14, 1121–1131 (2007).
4. B. W. Kesser, G. T. Hashisaki, J. R. Holt, Gene transfer in human vestibular epithelia and the
prospects for inner ear gene therapy. Laryngoscope 118, 821–831 (2008).
5. O. Akil, R. P. Seal, K. Burke, C. Wang, A. Alemi, M. During, R. H. Edwards, L. R. Lustig, Restoration
of hearing in the VGLUT3 knockout mouse using virally mediated gene therapy. Neuron 75,
283–293 (2012).
6. S.-I. Kitajiri, R. McNamara, T. Makishima, T. Husnain, A. U. Zafar, R. A. Kittles, Z. M. Ahmed,
T. B. Friedman, S. Riazuddin, A. J. Griffith, Identities, frequencies and origins of TMC1
mutations causing DFNB7/B11 deafness in Pakistan. Clin. Genet. 72, 546–550 (2007).
7. A. Sirmaci, D. Duman, H. Öztürkmen-Akay, S. Erbek, A. İncesulu, B. Öztürk-Hişmi, Z. S. Arici,
E. B. Yüksel-Konuk, S. Taşir-Yilmaz, S. Tokgöz-Yilmaz, F. B. Cengiz, İ. Aslan, M. Yildirim,
A. Hasanefendioğlu-Bayrak, A. Ayçiçek, İ. Yilmaz, S. Fitoz, F. Altin, H. Özdağ, M. Tekin,
Mutations in TMC1 contribute significantly to nonsyndromic autosomal recessive sensori-
neural hearing loss: A report of five novel mutations. Int. J. Pediatr. Otorhinolaryngol. 73,
699–705 (2009).
8. Y. Kawashima, K. Kurima, B. Pan, A. J. Griffith, J. R. Holt, Transmembrane channel-like (TMC)
genes are required for auditory and vestibular mechanosensation. Pflugers Arch. 467, 85–94
(2015).
9. H. Nakanishi, K. Kurima, Y. Kawashima, A. J. Griffith, Mutations of TMC1 cause deafness by
disrupting mechanoelectrical transduction. Auris Nasus Larynx 41, 399–408 (2014).
10. Y. Kawashima, G. S. G. Géléoc, K. Kurima, V. Labay, A. Lelli, Y. Asai, T. Makishima, D. K. Wu,
C. C. Della Santina, J. R. Holt, A. J. Griffith, Mechanotransduction in mouse inner ear hair
cells requires transmembrane channel-like genes. J. Clin. Invest. 121, 4796–4809 (2011).
11. B. Pan, G. S. G. Géléoc, Y. Asai, G. C. Horwitz, K. Kurima, K. Ishikawa, Y. Kawashima, A. J. Griffith,
J. R. Holt, TMC1 and TMC2 are components of the mechanotransduction channel in hair cells
of the mammalian inner ear. Neuron 79, 504–515 (2013).
12. M. Beurg, W. Xiong, B. Zhao, U. Müller, R. Fettiplace, Subunit determination of the con-
ductance of hair-cell mechanotransducer channels. Proc. Natl. Acad. Sci. U.S.A. 112, 1589–1594
(2015).
13. W. Marcotti, A. Erven, S. L. Johnson, K. P. Steel, C. J. Kros, Tmc1 is necessary for normal
functional maturation and survival of inner and outer hair cells in the mouse cochlea. J. Physiol.
574, 677–698 (2006).
14. S. Vreugde, A. Erven, C. J. Kros, W. Marcotti, H. Fuchs, K. Kurima, E. R. Wilcox, T. B. Friedman,
A. J. Griffith, R. Balling, M. Hrabé De Angelis, K. B. Avraham, K. P. Steel, Beethoven, a mouse
model for dominant, progressive hearing loss DFNA36. Nat. Genet. 30, 257–258 (2002).
15. Y. Zhao, D. Wang, L. Zong, F. Zhao, L. Guan, P. Zhang W. Shi, L. Lan, H. Wang, Q. Li, B. Han,
L. Yang, X. Jin, J. Wang, J. Wang, Q. Wang, A novel DFNA36 mutation in TMC1 orthologous
to the Beethoven (Bth) mouse associated with autosomal dominant hearing loss in a Chinese
family. PLOS One 9, e97064 (2014).
16. V.H. Hernandez, A. Gehrt, K. Reuter, Z. Jing, M. Jeschke, A. Mendoza Schulz, G. Hoch, M. Bartels,
G. Vogt, C. W. Garnham, H. Yawo, Y. Fukazawa, G. J. Augustine, E. Bamberg, S. Kügler, T. Salditt,
L. de Hoz, N. Strenzke, T. Moser, Optogenetic stimulation of the auditory pathway. J. Clin.
Invest. 124, 1114–1129 (2014).
17. S. Safieddine, R. J. Wenthold, SNARE complex at the ribbon synapses of cochlear hair cells:
Analysis of synaptic vesicle- and synaptic membrane-associated proteins. Eur. J. Neurosci.
11, 803–812 (1999).
18. J. C. Bedrosian, M. A. Gratton, J. V. Brigande, W. Tang, J. Landau, J. Bennett, In vivo delivery
of recombinant viruses to the fetal murine cochlea: Transduction characteristics and long-
term effects on auditory function. Mol. Ther. 14, 328–335 (2006).
19. L. A. Kilpatrick, Q. Li, J. Yang, J. C. Goddard, D. M. Fekete, H. Lang, Adeno-associated virus-
mediated gene delivery into the scala media of the normal and deafened adult mouse ear.
Gene Ther. 18, 569–578 (2011).
20. J. E. Gale, W. Marcotti, H. J. Kennedy, C. J. Kros, G. P. Richardson, FM1-43 dye behaves as a per-
meant blocker of the hair-cell mechanotransducer channel. J. Neurosci. 21, 7013–7025 (2001).
21. J. R. Meyers, R. B. MacDonald, A. Duggan, D. Lenzi, D. G. Standaert, J. T. Corwin, D. P. Corey,
Lighting up the senses: FM1-43 loading of sensory cells through nonselective ion chan-
nels. J. Neurosci. 23, 4054–4065 (2003).
22. G. S. Géléoc, J. R. Holt, Developmental acquisition of sensory transduction in hair cells of
the mouse inner ear. Nat. Neurosci. 10, 1019–1020 (2003).
23. A. J. Hudspeth, Integrating the active process of hair cells with cochlear function. Nat. Rev.
Neurosci. 15, 600–614 (2014).
24. M. C. Liberman, J. Gao, D. Z. Z. He, X. Wu, S. Jia, J. Zuo, Prestin is required for electromotility
of the outer hair cell and for the cochlear amplifier. Nature 419, 300–304 (2002).
25. S. F. Maison, X.-P. Liu, D. E. Vetter, R. A. Eatock, N. M. Nathanson, J. Wess, M. C. Liberman,
Muscarinic signaling in the cochlea: Presynaptic and postsynaptic effects on efferent
feedback and afferent excitability. J. Neurosci. 30, 6751–6762 (2010).
26. D. Grimm, M. A. Kay, J. A. Kleinschmidt, Helper virus-free, optically controllable, and two-
plasmid-based production of adeno-associated virus vectors of serotypes 1 to 6. Mol. Ther.
7, 839–850 (2003).
27. A. E. Stauffer, J. R. Holt, Sensory transduction and adaptation in inner and outer hair cells of
the mouse auditory system. J. Neurophysiol. 98, 3360–3369 (2007).
Acknowledgments: We thank M. C. Liberman for assistance with ABR recordings; Y. Shu and
Z.-Y. Chen for technical assistance with RWM injections; V. Padrun, F. Pidoux, and A. Aebi at
École Polytechnique Fédérale de Lausanne (EPFL) Brain Mind Institute for technical assistance
with AAV production; Behavior and Viral Cores at Boston Children’s Hospital [supported by
Boston Children’s Hospital Intellectual and Developmental Disabilities Research Center (BCH
IDDRC), P30 HD18655]. Funding: This work was supported by the Bertarelli Foundation, Kidz b
Kidz Foundation (Jessica and David Freier), and Program in Translational Neuroscience and
Neuroengineering. Author contributions: C.A. performed experiments, analyzed data, and
helped write the manuscript; C.R. designed and generated vectors; B.P. performed experiments
and analyzed data; Y.A. designed and generated vectors, performed experiments, and analyzed
data; H.A. performed experiments and analyzed data; E.C. analyzed data; B.L.S. designed and gen-
erated vectors and helped design experiments; P.A. helped design vectors and experiments; J.R.H.
conceived the study, designed experiments, analyzed data, and helped write the manuscript. All
authors critically reviewed and approved the manuscript. Competing interests: The authors de-
clare that they have no competing interests. Data and material availability: Plasmids for AAV
vector production are available via a material transfer agreement.
Submitted 13 December 2014
Accepted 17 June 2015
Published 8 July 2015
10.1126/scitranslmed.aab1996
Citation: C. Askew, C. Rochat, B. Pan, Y. Asai, H. Ahmed, E. Child, B. L. Schneider, P. Aebischer,
J. R. Holt, Tmc gene therapy restores auditory function in deaf mice. Sci. Transl. Med. 7,
295ra108 (2015).

DOI: 10.1126/scitranslmed.aab1996
, 295ra108 (2015);7Sci Transl Med
et al.Charles Askew
gene therapy restores auditory function in deaf miceTmc
Editor's Summary
implants and hearing aids.
will work long-term to maintain hearing recovery, perhaps supplementary existing technologies such as cochlear
mutations have been implicated in recessive prelingual deafness, so it is hoped that this gene therapeutic approach
TMC1and partially restore hearing, as determined by auditory brainstem responses and startle reflexes. More than 30
dominant human deafness, respectively. Both vectors were able to transduce inner hair cells of the mouse cochlea,
models representative of autosomal recessive and−−BeethovenorTmc1delivered to mice with mutations in
, was packaged in adeno-associated viral vectors andTmc2, or its ortholog,Tmc1like 1,−transmembrane channel
replace mutant genes associated with the mechanotransduction machinery of the inner ear. The gene encoding
Because genetics is a major cause of deafness, Askew and colleagues developed a therapeutic approach to
Can you hear me now?
/content/7/295/295ra108.full.html
can be found at:
and other services, including high-resolution figures,A complete electronic version of this article
/content/suppl/2015/07/06/7.295.295ra108.DC1.html
can be found in the online version of this article at:Supplementary Material
http://stm.sciencemag.org/content/scitransmed/2/21/21ra16.full.html
http://stm.sciencemag.org/content/scitransmed/6/233/233ra54.full.html
can be found online at:Related Resources for this article
http://www.sciencemag.org/about/permissions.dtl
in whole or in part can be found at:article
permission to reproduce thisof this article or about obtainingreprintsInformation about obtaining
is a registered trademark of AAAS.Science Translational Medicinerights reserved. The title
NW, Washington, DC 20005. Copyright 2015 by the American Association for the Advancement of Science; all
last week in December, by the American Association for the Advancement of Science, 1200 New York Avenue
(print ISSN 1946-6234; online ISSN 1946-6242) is published weekly, except theScience Translational Medicine

Tmc gene therapy restores auditory function in deaf mice

Recommandé

Recommandé

Contenu connexe

Tendances

Tendances (20)

En vedette

En vedette (8)

Similaire à Tmc gene therapy restores auditory function in deaf mice

Similaire à Tmc gene therapy restores auditory function in deaf mice (20)

Plus de José Luis Moreno Garvayo

Plus de José Luis Moreno Garvayo (20)

Dernier

Dernier (20)

Tmc gene therapy restores auditory function in deaf mice